DOI 10.20544/TTS2018.

P21
UDK 656.13.057:519.872(497.711)

EVALUATION OF ADAPTIVE AND FIXED TIME

TRAFFIC SIGNAL STRATEGIES: CASE STUDY OF
SKOPJE¹
Daniel Pavleski, Daniela Koltovska Nechoska
Faculty of Technical Sciences
St. Kliment Ohridski University, Bitola, Blvd 1st May, nn
Bitola, Republic of Macedonia
dhany_mk@hotmail.com

Edouard Ivanjko
Department of Intelligent Transportation Systems,
Faculty of Transport and Traffic Sciences,University of Zagreb,Vukelićeva 4
Zagreb, Republic of Croatia

Abstract

Signalized intersections and belonging signal plans are powerful means

for effective urban traffic management. Numerous strategies have evolved
over the time, from fixed time to intelligent signal plan control strategies that
respond and react to traffic conditions in real time. This paper presents an
evaluation and comparison of the impact of fixed time and UTOPIA
adaptive traffic signal control strategies for the chosen urban corridor in the
wider center area of the City of Skopje, Macedonia. A topic that still requires
more attention from the research community. This paper presents a
continuation of the author’s previous research where a framework to test
UTOPIA adaptive control using the microscopic simulator VISSIM was
described. Obtained new results include the influence of fixed time signal
plans which are compared with UTOPIA’s results and analyzed in details
using the following performance measures: delay, queue length, travel time,
intersection level of service, number of stops and vehicle throughput.

Keywords - signal plan control strategies; intelligent transportation

systems; UTOPIA; microscopic simulation; urban corridor

¹ professional paper


 INTRODUCTION
Intelligent transport systems (ITS) that embed advanced technologies in
transport are being accepted by transport authorities around the world as a
viable alternative to reliance on building more roads to reduce congestion
since many years [1]. In many experiments around the world, it has been
shown that Advanced Traffic Management Systems (ATMS) can improve
traffic quality by reducing traffic congestion, delays, and saving traveling
time [2, 3]. In large cities, such systems are implemented in their respective
Urban Traffic Control (UTC) centers.
Traffic signal control is a core element for efficient operation of the urban
network and crucial part of ITS [1]. Traffic signal control strategies are the
most common means used for managing the traffic flows at different types
of intersections in the cities. Today many cities in the world still implement
fixed time control strategies. The disadvantage of these strategies is that they
do not respond to everyday fluctuations in traffic patterns. Adaptive traffic
control strategies are designed to overcome the limitations of fixed time
control and they respond to fluctuations in everyday traffic patterns by
adjusting signal timing in accordance with current traffic demand and can
reduce traffic congestion, delays, and travel time. As an example, systems
like SCATS, SCOOT, RHODES, ImFlow, and UTOPIA (Urban Traffic
Optimization by Integrated Automation) can be mentioned [2]. Each of the
mentioned systems has a unique characteristic. For example, UTOPIA
/SPOT (System for Priority and Optimization of Traffic), which is used in
the city of Skopje, is a hierarchical decentralized traffic signal control
strategy [4]. It is also used in many cities in Italy as well as Netherlands,
USA, Norway, Finland, and Denmark. It minimizes the total time lost by
private vehicles during their trips, subject to the constraint that public
vehicles to be prioritized shall not be stopped at signalized intersections [2].
The effectiveness of the mentioned adaptive traffic control systems has to
be evaluated or at least estimated on the chosen urban traffic network before
implementation. As obtained in [2], adaptive traffic control cannot obtain an
optimal solution in all traffic scenarios. Especially in a case of a corridor or a
larger complex urban traffic network. To find which traffic control approach
is optimal and to tune the controller, simulations are used. This paper
presents an evaluation and comparison of the impact of fixed time and
UTOPIA adaptive traffic signal control strategies for the chosen urban
corridor in the wider center area of the City of Skopje, Macedonia. It is a
continuation of the author’s previous research where a framework to test
UTOPIA adaptive control using the microscopic simulator VISSIM was
described, implemented and tested [5]. Obtained new results about the
influence of fixed time signal plans are compared with UTOPIA’s results


and analyzed in details in the continuation of this paper using the following
performance measures: delay, queue length, travel time, intersection level of
service, number of stops and vehicle throughput.
This paper is organized as follows. The second section describes the
chosen performance measures for urban traffic control. Basic description of
booth evaluated traffic signal strategies is presented in the third section. The
fourth section presents the implemented simulation setup. The fifth section
presents obtained results and discussion about them. Conclusion and future
work description ends the paper.

PERFORMANCE MEASURES FOR URBAN TRAFFIC CONTROL

To evaluate the performance of different traffic signal control strategies
appropriate measures of effectiveness (MoE) should be determined and
discussed. The available literature describes a variety of MoEs that can be
used for this task in order to cope with different site-specific issues and with
different performance metrics that can be applied for an urban environment
[6, 7]. One of the site-specific issues can be the influence of traffic from side
streets [8]. In this use case, a part of the urban network with several main
streets is analyzed so MoEs for side street traffic like side street delay and
delays for major turning approaches are not examined.
To give a clear description and to understand traffic flow conditions at an
individual intersection under fixed time and adaptive traffic control
strategies the following performance measures are being applied: delay,
level of service (LoS), average queue length, max queue length, numbers of
stops and vehicle throughput. MoEs defined in previous author’s work [5]
are augmented with vehicle throughput to better evaluate the performance of
fixed traffic signal strategies and UTOPIA. The reasons for determining
these parameters are as follows. Delay and LoS play a primary role in
determining individual intersection performance. LoS can be used to
understand the quality of traffic conditions on a particular intersection and
delay exposes the difference between free-flow and congested traffic
conditions. Frequent stops due to congestion are a typical characteristic of
urban traffic. One of the reasons for this is the presence of signalized
intersections. Therefore, the information of queue length and number of
stops must be included as a performance measure also. The vehicle
throughput can provide useful information about the maximum number of
vehicles which can be discharged during the time as a measure of the
comparative effectiveness of the two evaluated traffic control strategies. In
addition, for an in-depth analysis of the arterial section in the analyzed urban
traffic network, travel time, delay and number of stops were obtained.


VISSIM microscopic traffic simulator software was used to obtain all these
mentioned MoEs [9].

TRAFFIC SIGNAL STRATEGIES FOR EVALUATION

As mentioned, the primary emphasis of this paper is to evaluate and
compare the impact of fixed time and UTOPIA adaptive traffic signal
control strategies. A brief review regarding these strategies is presented
below as a background for further analysis.

Fixed time control

Conventional traffic signal control approaches are either fix-timed or
vehicle actuated [2]. Fixed time control uses predetermined green and cycle
time that has a fixed duration, regardless of changes in traffic volumes
during the day. This type of control gives the most of the green time to the
heaviest traffic movement based on historical data. Some fixed-time systems
use different preset time intervals for the morning peak hour, evening peak
hour, and other busy periods. For this reason, such an approach cannot cope
with an unexpected change in traffic demand. To improve this, vehicle
actuated traffic signals are implemented. Conventional vehicle actuated
signals in principle work with green-demand and green-extension based on
vehicle detection, within pre-set limits, as long as the headways in the traffic
stream are short enough to be accurately measured. The cycle time is
variable, and adjacent signalized intersections can therefore only be
coordinated in special cases [2].
In Macedonia, signal control strategies are signal group based and they
operate under isolated and coordinated fixed time control. Exclusion is the
city of Skopje where the UTOPIA adaptive traffic signal control has been
implemented in order to reduce traffic congestion. Before UTOPIA was
implemented, the city of Skopje also used isolated and coordinated fixed
time control.

Adaptive control system UTOPIA

As mentioned, UTOPIA is an adaptive traffic control system designed to
optimize traffic flows and give selective priority to public transport without
sacrificing travel times for private traffic [2, 3]. UTOPIA is an innovation in
Urban Traffic Control (UTC) i.e. it is a hierarchical, adaptive, distributed
and open traffic control system [4]. Hierarchical means that it has goal
related coordination and includes cooperative control. Adaptive means that
the system monitors the traffic on-line and optimizes the signal plans to
ensure fast responses to changes in traffic demand. A distributed system


comprises decomposition, looking ahead, strong interaction and looking for
terminal costs. The overall network optimization is decomposed into
problems of coordinated junctions solved by the intersection units (SPOT) in
collaboration with the central system. In Torino UTOPIA/SPOT has resulted
in reduced travel times of 2-7% for public transport (PT) vehicles and 10%
for cars [10]. Field tests of UTOPIA/SPOT in Skopje have shown 20% travel
time reduction in peak hours [11].

SIMULATION SETUP
For evaluation and comparison of the impact of fixed time and UTOPIA
adaptive signal control strategies in this research, the microscopic simulator
VISSIM was applied. VISSIM is a microscopic, time step and behavior
based simulation model developed to analyze the full range of functionally
classified roadways and public transportation operations [4, 9]. VISSIM can
model integrated roadway networks found in a typical corridor as well as
various modes consisting of general purpose traffic, buses, light rail, heavy
rail, trucks, pedestrians, and bicyclists. This section presents the simulation
framework for strategies evaluation, chosen urban network for testing and
simulation parameters.

Simulation framework
A. Fixed time control
The steps for developing the simulation model were the following: (i)
creation of the road network using background image with links and
connectors; (ii) definition of traffic compositions observed from the field;
(iii) definition of speed distributions; and (iv) definition of routing which
was done based on the observed travel pattern of vehicles entering and
leaving the study area. The real-time traffic volumes were obtained by
Traffic Management Control Centre (TMCC) in Skopje of every signalized
intersection. Traffic signal controls were defined in VISSIM using the
internal fixed time controller. Signal heads were placed in the model to
represent available stop lanes. Signal timings of fixed controllers have been
obtained from the detailed TMCC traffic design projects. Data collection
points were added to the simulation model to correspond with actual data
collection points in the real world urban network.

B. VISSIM-UTOPIA
The adaptive traffic signal control system UTOPIA and microscopic
simulator VISSIM were connected using the UTOPIA-VISSIM Adapter
(UVA). The UTOPIA system is used as” black box” control unit connected
to the traffic network simulated in VISSIM. The method used in this


research was ”software-in-the-loop simulation environment”. More
information regarding the simulation framework and detailed description of
testing of the UTOPIA and VISSIM connection can be found in [5].

Chosen urban road network

In order to evaluate the impact of traffic signal control strategies, a
simulation model was created in VISSIM for the congested corridor of 7
signalized intersections in the wider center area of the City of Skopje,
Macedonia as illustrated in Fig. 1.

Fig. 1. Spatial configuration of the study area [5]

The traffic demand is higher in the morning peak hour (07:20-08:20) for
the chosen urban network because most of the trips are realized from the east
part of the city to the city center. Therefore, this peak hour is used for
evaluation in this paper. Total traffic flows in the morning peak hour for all
examined intersections respectively are presented in Table 1. The highest
traffic flows occur at the intersection I6.

Table 1. Total traffic flows at the analyzed intersections [5]

Intersection Total traffic flow (veh/h)
I1 3,614
I2 4,055
I3 1,913
I4 2,147
I5 3,907
I6 5,110
I7 1,477


 Simulation parameters
To create a realistic simulation model, traffic data were obtained from
loop detectors mounted in the study area as mentioned. All bus stops located
in the study area have been included in the VISSIM model also. Bus routes
of public transport are modeled using officially published timetables. In this
phase of research bicycle and pedestrian traffic is not analyzed. Five
simulations were performed for booth traffic signal control strategies and
average values were computed to obtain realistic results. The simulation time
covers the morning peak hour in a typical working day. A 900-second warm-
up period was used before data collection for analysis started. This warm-up
period is needed to fill the empty road network with vehicles and create a
realistic traffic situation for simulation testing.

RESULTS AND DISCUSSION

Obtained averaged simulations results for booth strategies are presented
in this section. Tables 2 and 3 contain the simulation results for MoEs
obtained from the evaluation of fixed time traffic signal strategy. Table 2
presents LoS, delay, average queue length, max queue length, number of
stops and the vehicle throughput obtained for every single intersection in the
study area. Additionally, Table 3 presents the travel time, delay, number of
stops, and number of vehicles obtained for the arterial section between
intersections I1 and I6.

Table 2. Obtained MoE for all intersections under fixed time traffic signal strategy
MoE I1 I2 I3 I4 I5 I6 I7
LoS D C B E A E D
Delay (s) 54.11 27.77 18.64 55.68 7.50 68.98 53.75
QLen (m) 75.01 19.98 7.52 23.46 4.26 56.78 56.47
QLenMax
510.1 161.33 83.65 127.91 77.17 242.08 240.59
(m)
Stops 0.92 0.64 0.56 0.96 0.30 1.46 1.34
Vehicle
throughput 2,830 3,314 1,485 1,727 3,303 4,587 997
(veh/h)

Table 3. Obtained MoE for arterial sections under fixed time traffic signal strategy
MoE I1 → I6 (L = 1,080 m) I6 → I1 (L = 679 m)
Travel Time (s) 442.80 70.06
Delay (s) 368.28 22.87
Stops 8.29 1.12


 As shown in Table 2, the majority of the intersections operate at an
acceptable LOS of B, C, or D during the morning peak hour. Only two
intersections (I4 and I6) operate at unacceptable LoS E during the peak hour.
A significant delay is occurring at intersections I4 and I6, the ones with the
worst LoS. Regarding other MoEs, the longest average queue length and the
maximal queue length is appearing at intersection I1. The worst number of
stops is appearing at intersection I6. The maximum number of vehicles
which can be discharged during the time appears at I6. From Table 3 one can
conclude that travel time, delay, and the number of stops is higher in
direction I1 → I6 during the morning peak hour.

Table 4. Obtained MoE for all intersections under UTOPIA

MoE I1 I2 I3 I4 I5 I6 I7
LoS D D B D A E B
Delay (s) 52.08 43.47 15.89 38.10 7.23 65.21 16.43
QLen (m) 64.84 39.92 6.58 17.70 3.76 43.87 8.86
QLenMax
409.84 213.59 87.95 135.62 108.55 273.03 55.32
(m)
Stops 1.50 1.24 0.57 1.14 0.34 1.38 0.50
Vehicle
throughput 3,163 3,712 1,861 2,065 3,593 4,795 1,421
(veh/h)

Table 5. Obtained results for arterial sections under UTOPIA

MoE I1 → I6 (L = 1,080 m) I6 → I1 (L = 679 m)
Travel Time (s) 366.72 97.02
Delay (s) 292.68 50.04
Stops 9.53 1.39

Tables 4 and 5 contain the simulation results for MoEs obtained from the
evaluation of UTOPIA adaptive traffic signal strategy. As shown in Table 4,
most of the intersections operate at an acceptable LoS of B, C, or D during
the morning peak hour. Only one intersection (I2) operate at unacceptable
LoS E during the peak hour. A significant delay is occurring at intersection
I6, the one with the worst LoS. Regarding other MoEs, the longest average
queue length and the maximal queue length is appearing at intersection I1.
The worst number of stops is appearing at intersection I1 also. The
maximum number of vehicles which can be discharged during the time
appears at intersection I6. From Table 5 one can conclude that travel time,
delay, and the number of stops is higher in direction I1 → I6 during the
morning peak hour.


 Adaptive traffic signal control strategy shows better results regarding
LoS, average queue length, maximal queue length, and vehicle throughput.
Only the number of stops increases under this type of control. From the
analysis of obtained results for arterial sections, adaptive control strategy
shows an overall decrease of travel time and delay. A small increase can be
noticed in the number of stops. Field tests show even better results because
in the congested period the traffic operator includes a preemption strategy
that was not investigated in this phase of research.

CONCLUSION AND FUTURE WORK

The results of the evaluation showed that the performance of the UTOPIA
adaptive traffic signal control provided better LoS, average queue length,
maximal queue length and vehicle throughput across the intersections versus
fixed time control. A small increase can be noticed in the number of stops
under adaptive control strategy. The overall travel time is decreased in the
examined arterial corridor showing the influence of adaptation to different
traffic demand regarding travel direction. Namely, UTOPIA decreases the
travel time in the congestion direction and at the same time the travel time in
the opposite direction increases. Thereby, the benefits gained in the
congested direction are larger than the drawback in the opposite direction.
In the recent decade, algorithms from the domain of artificial intelligence
are being used for traffic control. They have the ability to accumulate and
use knowledge, set a problem, learn, process, conclude, solve the problem
and exchange knowledge with other systems. Future work on this topic will
include comparison and testing of self-learning urban traffic control
strategies versus UTOPIA adaptive control.

Acknowledgement
The authors would like to thank the companies PTV Group and SWARCO
MIZAR S.r.l., the Traffic management and control center of the City of
Skopje, and the Faculty of Transport and Traffic Sciences, University of
Zagreb for supporting the work published in this paper.

REFERENCES
[1] D. Zhao, Y. Dai, Z. Zhang, “Computational Intelligence in Urban Traffic
Signal Control: A Survey,” IEEE Transactions on Systems, Man, and
Cybernetics-Part C: Applications and Reviews, Vol. 42, No. 4, July 2012,
pp. 485-494
[2] J. Wahlstedt, “Evaluation of the two self-optimising traffic signal
systems Utopia/Spot and ImFlow, and comparison with existing signal
control in Stockholm, Sweden,” in Proceedings of the 16th International


IEEE Annual Conference on Intelligent Transportation Systems (ITSC
2013), Hague, Netherlands, October 6-9, 2013. pp. 1541–1546.
[3] S. Samadi, A. P. Rad, F. M. Kazemi, H. Jafarian, „Performance
Evaluation of Intelligent Adaptive Traffic Control Systems: A Case Study,“
Journal of Transportation Technologies, 2012, 2, pp. 248-259, DOI:
jtts.2012.23027
[4] D. Taranto, “UTOPIA Urban Traffic Control Overview,” presentation,
Faculty of Technical Sciences, Bitola, Macedonia, December 2012.
[5] D. Pavleski, D. Nechoska Koltovska, E. Ivanjko, “Evaluation of
Adaptive Traffic Control System UTOPIA using Microscopic Simulation,”
Proceedings of 59th International Symposium ELMAR-2017, Special
session Intelligent Transport Systems, Zadar, Croatia, 18.-20.09.2017. pp.
17-20
[6] X. Ban, J. Wojotowicz, and W. Li, “Decision-Making Tool for Applying
Adaptive Traffic Control Systems,” Final report C-13-04, No. 16-12,
Rensselaer Polytechnic Institute, Troy, NY, USA, March 2016.
[7] M. Hunter, S. Wu, H. Kim, and W. Suh, “A probe-vehicle-based
evaluation of adaptive traffic signal control,” IEEE Transactions on
Intelligent Transportation Systems, vol. 13, pp. 704–713, June 2012.
[8] M. Hunter, S. Wu, and H. Kim, “Evaluation of adaptive traffic signal
control: Case study of Cobb County SCATS,” in Proceedings of the 17th
ITS World Congress, (Busan, Korea), October 25-29, 2010.
[9] PTV Group, “PTV Vissim simulation software.” http://vision-
traffic.ptvgroup.com. [accessed 05 December 2015.].
[10] J. Wahlstedt, “Impacts of Bus Priority in Coordinated Traffic Signals,”
Proceedings of the 6th International Symposium on Highway Capacity and
Quality of Service Stockholm, Sweden, June 28 – July 1, 2011. pp. 578–587
[11] SWARCO MIZAR S.p.A., ATM Skopje Preliminary Before/After
Study, Turin, 2014.